Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 38

Critical Review of

Analytical and Numerical


Modeling of Pile
Foundations Subjected to
Earthquake Loading
Mohammad Reza Falamarz-Sheikhabadi
Candidacy Exam
Fall 2012

Drexel University, Civil Engineering Department 1

Contents

Pile foundation definition


Different methods for pile modeling
Winkler method
Beam on nonlinear Winkler foundation (BNWF)
Finite element method

Absorbing boundary conditions


Viscous damping boundary model
Perfectly matched layer

Discussion on BNWF
Conclusion
References
Drexel University, Civil Engineering Department 2

Pile Foundation
Definition

Drexel University, Civil Engineering Department

Pile Foundations
Pile usually represents a
slender
structural
element that is driven
into
the
ground.
However, a pile is often
used as a generic term to
represent all types of
deep foundations.
Here, I talk about normal
size piles (250-600 mm)

Ref. 1

Drexel University, Civil Engineering Department 4

Pile Foundations
Different types of pile
foundations to carry
vertical, horizontal or
inclined loads from
superstructure.

Ref. 2

Drexel University, Civil Engineering Department 5

Individual Pile and Pile


Group
Pile foundations may
consists of an individual
pile or a group of piles.
Although a pile group is
composed of a number
of individual piles, the
behavior of a pile group
is not equivalent to the
sum of all the piles as if
they
were
separate
individual piles
Ref. 2

Drexel University, Civil Engineering Department 6

Individual Pile and Pile


Group
The behavior of a pile
group is more complex
than an individual pile
because of the effect
of the combination of
piles,
interactions
between the piles in
the group, and the
effect of the pile cap.

Ref. 2

Drexel University, Civil Engineering Department 7

Pile Modeling Methods

Drexel University, Civil Engineering Department

Pile Modeling

Continuum method
Winkler method
Beam on nonlinear Winkler foundation
Finite element method

Ref. 3

Drexel University, Civil Engineering Department 9

Winkler method
The Winkler approach is the
oldest method for estimating
pile deflections and bending
moments.
The
approach
model the soil as a series of
unconnected linear springs
with stiffness, Es, expressed in
units of force per length.

Ref. 2

Drexel University, Civil Engineering Department 10

Winkler method
Limitations:
1. The method ignores the
nonlinear characteristics of
soil.
2. The axial load effects are
ignored.
3. The soil model used in the
technique is discontinuous.

Ref. 2

Drexel University, Civil Engineering Department 11

Beam on nonlinear Winkler


foundation (BNWF)
The
py
method
is
versatile and can be used
to
solve
problems
including different soil
types,
layered
soils,
nonlinear soil behavior,
different pile materials,
cross
sections,
and
different
pile
head
connection conditions.

Ref. 4

Confining pressure

Drexel University, Civil Engineering Department 12

Beam on nonlinear Winkler


foundation
Considering
that
both
piles and soil can behave
in a nonlinear manner
during extreme events,
the use of p-y methods for
defining
the
lateral
stiffness of pile-soil model
for
seismic
analysis
(secant stiffness as a
function
of
pile
deformation) has been
used since the seventies.
Ref. 5
Why sometimes I have used p-y elements and sometimes p-y springs?

Drexel University, Civil Engineering Department 13

Beam on nonlinear Winkler


foundation
In this method, the reaction of soil
surrounding the pile is modeled as
localized springs: a series of springs
along the shaft (the t-z curves) and
the spring attached to the tip or
bottom of a pile (the Q-z curve). The
load transfer or unit friction force
along the shaft is shown by t, Q is
the tip resistance of the pile in
compression, and z is the settlement
(for static analysis) or vertical
deformation (for dynamic analysis)
of soil at the location of a spring.
Ref. 1

Drexel University, Civil Engineering Department 14

BNWF and Pile group


Since analyzing the
pile group supported
structures
under
lateral loading needs
including the effects
of rocking motions, in
the
BNWF,
all
localized
nonlinear
springs, namely, p-y,
t-z and Q-z should be
considered

Ref. 6

Drexel University, Civil Engineering Department 15

BNWF and Pile group


The
bending
moments
for
the
corner piles should be
increased for closely
spaced piles
Side-byside
spacing

Corner pile
moment
modificatio
n factor

3D

2D

1.2

1D

1.6
Ref. 8
Ref. 2

Drexel University, Civil Engineering Department 16

P-multiplier
A popular method to
account for shadowing
effect is to incorporate
p-multipliers into the p-y
method of analysis. The
p-multiplier
values
depend on pile position
within the group and pile
spacing.

Ref. 7

Drexel University, Civil Engineering Department 17

P-multiplier
Factors influencing p-multiplier:
pile spacing
group arrangement
group size
pile head fixity
soil type and density

Ref. 8

Drexel University, Civil Engineering Department 18

Beam on nonlinear Winkler


foundation
Limitations:
1. The p-y method is based on pseudo-static loading for while
lateral forces from the upper structure are only applied,
2. The accuracy of the py method depends on the number of
tests and the variety of tested parameters, such as geometry
and stiffness of pile, layers of soil, strength and stiffness of
soil, and loading conditions.
3. The effects of pile diameter have not been considered in the
primary relations of p-y curves.
4. The pile cap effects are usually ignored and cap pile are
considered rigid.
5. For pile groups, using p-multiplier method oversimplifies the
problem.

Drexel University, Civil Engineering Department 19

Finite element method


The FE method has the ability of permitting
to account for soil nonlinearity by applying
appropriate constitutive models, such as the
Drucker-Prager, Cam-Clay or Mohr Coulomb
formulation, and to use gap-elements to
model possible pile soil separation.

Ref. 9
Too difficult and
time consuming!

Drexel University, Civil Engineering Department 20

Finite element method


Limitations:
1. The cost of the specialized software.
2. The
time
consuming
model
generation.
3. The time required for the non-linear
analysis.
4. The difficulty in the interpretation of
the result in terms of common pile
(beam) variables.
5. The uncertainties associated with
soil non-linear modeling in 3D.
Ref. 9

Drexel University, Civil Engineering Department 21

Absorbing Boundary
Conditions

Drexel University, Civil Engineering Department 22

Absorbing boundary conditions

Bounded medium

Semi-infinite medium

The boundaries absorb the waves should be transmitted in semi-infinite medium.

Bounded
medium

Drexel University, Civil Engineering Department 23

Absorbing boundary conditions

Viscous damping boundary method


Perfectly matched layer
Infinite elements
Consistent dashpot, spring and mass method
And other numerical methods

Ref. 3

Drexel University, Civil Engineering Department 24

Viscous damping boundary


method
The simplest local ABC is
the
classical
normal
impedance or standard
viscous
boundary.
Its
performance is known to
deteriorate as the position
approaches the source of
perturbation especially in
the low-frequency limit.

This method has been modified in different manners!

Ref. 10

Drexel University, Civil Engineering Department 25

Viscous damping boundary


method
1. It can be only used for
dynamic analyses.
2. The method is able to
absorb only primary and
secondary waves under an
angle of incidence of 90o.
3. The
interested
medium
should be elastic and linear.

Ref. 10

Drexel University, Civil Engineering Department 26

Perfectly matched layer


The PML is only reflectionless
when
the
exact
wave
equation
is
solved.
In
practical applications, when
the approximate methods
like
finite-difference-timedomain (FDTD) or FE are
applied for modeling, the
analytical perfection of the
PML is no longer valid and so
user should consider this
point in the analysis.

!!!Soil=Elastic Medium!!!
Ref. 12

Drexel University, Civil Engineering Department 27

Perfectly matched layer


A newly discovered silent
boundary is the perfectly
matched layer method, first
introduced for the use of
electromagnetic waves. The
concept has been designed
designed
to
absorb
thoroughly
any
incident
wave without reflection, for
any incident angle and at
any
frequency
before
discretization.
Ref. 11

Drexel University, Civil Engineering Department 28

Perfectly matched layer


1. When the approximate methods
like finite-difference-time-domain
(FDTD) or FE are applied for
modeling, the analytical perfection
of the PML is no longer valid and
so user should consider this point
in the analysis.
2. Interested region should be elastic
(such an assumption for soft or
saturated soil condition and largescale structures is absolutely
unacceptable).
3. Its performance depends on the
type of seismic waves.

Ref. 11

Drexel University, Civil Engineering Department 29

Discussion on BNWF

Drexel University, Civil Engineering Department 30

Discussion

In spite of the fact that the seismic responses of structures are


resultant of the combined action of at least three translational
components of ground motion (ignoring two rocking and one
torsional earthquake components), the coupling effects of these
components in the modeling of pile foundations are commonly
ignored in the analytical and numerical studies.

Drexel University, Civil Engineering Department 31

Earthquake Components
50

40

15

40

30

30

20

20

10

aT (t) cm/s/s

aV(t) cm/s/s

10

0
-5
-10
-15
-20

20

40

60

80

10
0
-10

0
-10
-20

-20

-30

-30

-40

-40

100

aR(t) cm/s/s

20

20

40

60

80

-50

100

20

40

20

40

60

80

60

80

100

Translational components
15

20

15

15

0
-5
-10
-15

a R(t) mrad/s/s

a T (t) mrad/s/s

a V(t) mrad/s/s

10

10

10

-5

20

40

60

80

100

-10

0
-5
-10

-20
-25

20

40

60

80

100

-15

100

Rotational components
Drexel University, Civil Engineering Department 32

Earthquake Components
8000

8000

4000

7000

7000

6000

6000

5000

5000

3500

2500

A T ( )

AV()

3000

2000

AR()

4500

4000

4000

3000

3000

1000

2000

2000

500

1000

1000

1500

0
0

10

15

20

(Hz)

25

30

35

10

15

20
(Hz)

25

30

35

10

15

20

(Hz)

25

30

35

4000

1800

3500

1600

3000

1400

2500

1200

2000
1500

2000

1000
800
600

1000

400

500
0

2500

A R()

A T ()

A V()

Translational components

1500

1000

500

200
0

10

15

20

(Hz)

25

30

35

10

15

20

(Hz)

25

30

35

Rotational components

10

15

20

(Hz)

25

30

35

Drexel University, Civil Engineering Department 33

Discussion

Pile caps are often considered rigid and their effects in lateral
resistance of pile group are considered ignorable. Both of these
assumptions may cause an unknown error in estimation of actual
behavior of pile foundations.
Based on the size and configuration of pile group, spatial variation of
strong ground motions, uncertainty in distribution of mass and
stiffness of piles, unequal distribution stress on piles (shadowing
effects), probable damages and torsional earthquake component; a
considerable torsional moment may induce in the pile caps. This
effect is usually ignored in the typical two-dimensional analyses.
Although there are many recorded data on the ground surface due to
various earthquakes in different site conditions, it is not the case for
data recorded underground surface. Therefore, more data are
required to estimate the seismic loading of deep piles. It should be
pointed out that soil is more anisotropic and non-homogenous in
depth.

Drexel University, Civil Engineering Department 34

Conclusion

Drexel University, Civil Engineering Department 35

Conclusion
Based on the previous research on soil-pile interaction,
although it seems that the p-y method can be considered
as an efficient technique applicable to many practical
applications, it still needs some modifications and
developments in order to give more reliable results in the
soil-pile interaction analyses.

Drexel University, Civil Engineering Department 36

References
1.

Chen, W. F, Duan, L, Bridge Engineering Handbook, CRC press LLC (2000).

2.

Murthy, V. N. S, Geotechnical Engineering: Principles and practices of soil mechanics and


foundation engineering, Marcel Dekker, Inc.

3.

Shin, H, Arduino, P, Kramer, S. L, Mackie, K, Seismic response of a typical highway bridge in


liquefiable soil.

4.

Gazetas, G, Dobry, R, Simple Radiation Damping Model for Piles and Footings, Journal of
Engineering Mechanics ASCE, Vol. 110 (1984) 937-956.

5.

Boulanger, R. W, Curras, C. J, Kutter, B. L, Wilson, D. W, Abghari, A, Seismic soil-pile-structure


interaction experiments and analyses, Journal of Geotechnical and Geoenvironmental
Engineering, Vol. 125 (1999) 750-759.

6.

Curras, C. J, Boulanger, R. W, Kutter , B. L, Wilson, D. W, Dynamic experimental and analysis


of a pile-group-supported structure, Journal of Geotechnical and Geoenvironmental
Engineering, Vol. 127 (2001) 585-596.

7.

Comodromos, E. M, Papadopoulou, M. C, Explicit extension of the p-y method to pile groups


in cohesive soils, Computers and Geotechnics, Vol. 47 (2013) 28-41.

8.

Mokwa, R. L, Investigation of the resistance of pile cap to lateral loading, PhD thesis,
Virginia Polytechnic Institute and State University (1999).

9.

Yang, Z, Jeremic, B, Numerical analysis of pile behavior under lateral loads in layered elasticplastic soils, International Journal for Numerical Methods in Geomechanics, Vol. 2 (2002) 131.

10.

Kramer, Steven L, Geotechnical earthquake engineering, Prentice Hall, Inc., Upper Saddle
River, New Jersey, (1996).

11.

Basu, U, Chopra, A. K, Perfectly matched layers for time-harmonics elastodynamics of


unbounded domains: theory and finite-element implementation, Computer methods in
applied mechanics and engineering, 192 (2003) 1337-1375.

12.

Hasting, F. D, Schneider, J. B, Broschat, S. L, Application of the perfectly matched layer (PML)


absorbing boundary condition to elastic wave propagation, Journal of Acoustic Society of
America, 100 (1996) 3061-3069.

Drexel University, Civil Engineering Department 37

Thank you

38

You might also like